Neural Net Model for the Fluidized-bed Spray Granulation and Its Applications

A BP neural net model is constructed to map the relationship of process variables and physical properties of products in fluidized bed spray granulation.The structure and parameters of BP neural network are optimized,by using the methods of orthogonal experiments, intercross evaluating and training, normalization of the network training set and principal component analysis.Simulating outputs of the BP network are very close to the results of experiments and have wide adaptability. The BP network can be used conveniently to carry out various quantificational calculations. For example, the average particle size under specified process variables can be estimated, or appropriate process variables be determined according to given production target.